Simulation of methane emissions from rice paddies and tropical
wetlands in CLM4
Cornell: Lei Meng, Peter Hess,Natalie Mahowald, and Joseph Yavitt
In collaboration with Zack Subin and Bill Riley at Lawrence Berkeley National Lab
Outline
• Introduction
• Two new features in the methane model
• Model simulations against observations
• Conclusions
Introduction
• Methane emission from rice paddies
• Methane emission from wetlands
• Rice paddies + wetlands~200 Tg /yr, ~40% of global methane budget
• Large uncertainties in global methane budget
(Wuebbles &Hayhoe 2002) http://www.mccullagh.org/db9/vietnam/vietnam-rice-paddy-fields.jpg
Total anthropogenic sources of methane are about 358 Tg /year (IPCC 2007)
Natural sources: 145~260 Tg /yr; Wetlands: 100~231 Tg/yr
Model description (1)
• We used the methane model developed by Riley et al at LBNL.
)redox(*)pH(*)WTP(*)T(*HRP 4 ffffch =
Soil temperature
Water table position
Soil pH Redox potential
CLM-CN heterotrophic respiration
Model description (2)
unsatchsatchch FF _4_44 *)finundated1(*finundatedF −+=
Net CH4 flux at each grid cell
Fractional inundation at each grid cell
Net CH4 flux from saturated portion
Net CH4 flux from unsaturated portion
finundated1-finundated
Fch4_sat
Fch4_unsat
Distribution of Rice paddies and wetlands
• We forced the model with external fractional inundation and rice paddy fraction in order to remove potential errors associated with CLM hydrology.
(Matthews et al. 1991, Prigent et al. 2007)
pH dependence of methane production
6.8*7727.2*2335.0 2
10)f( −+−= pHpHpH
R =0.66
Field and laboratory datasets
(Data from Dunfield et al. 1993,Soil Biol. Biochem)
Advantage:Allow for methane production from bogs which often has low pH values
Impact of redox potential on methane production
Assumptions: 1. Newly inundated land will not produce methane initially, because of the availability of other electron acceptors (O2,SO4
-2, Fe3+, etc)2. As the other electron acceptors are consumed, the inundated fraction which will produce methane grows. The fraction with other electron acceptors decays with an e-folding time scale of 30 days. (chosen to match data/understanding)3. Redox potential will reduce methane production.
Simulations of tropical wetland (1)
PanamapH =6
IndonesiapH =4.0
Rice paddy simulation (1)
California 1982
Assumption: 1. Rice paddy fields are fully unsaturated before flooding and transplanting and are continuously flooded until harvest2. Under balanced condition, soil C will be degraded to 50% CO2 and 50% CH4
California 1983
Japan 1993Japan 1991
Rice paddy simulations (2)
ItalyTexas
Nanjing, China Chengdu, China
Global simulation-Rice paddy
• Global average: 106 Tg/year (Others:23 Tg/year -120 Tg/year)
-- assume continuous flooding (no drainage during growing seasons)
--overestimate methane emissions from Asian rice paddies
Global simulation-Wetlands
Our model Other models
Global 133 100-231
Tropical (20N-30S) 82 66- 88+
Units: Tg /year
Conclusions
• Two new features (pH and redox potential) have been added into the methane model.
• Preliminary results suggest that the improved methane model does a reasonable job in simulating methane emissions from tropical wetlands and rice paddy fields
Impact of Redox potential on methane production (2)
fredox(t) = finundated(t) – finundated(t-1) -newly inundated land+ fredox(t-1)*(1-Δt/tau) -30 days decay
tautfredox
dtfinundatedd
dtfredoxd )1()()( −
−=
fredox is the fraction of gridcell with other species (such as O2, SO4-2, Fe+3) to consume
finundated is the original fractional inundation of gridcell, finundated_adj is the adjusted fractional inundation, tau is the delay time (30 days) for other species, t is the current time step and t-1 is the previous time step
unsatchsatchch FF _4_44 *)_adjfinundated1(*_adjfinundatedF −+=
fredoxfinundated_adjfinundated −=
Similarly, the redox potential can also delay the methane production in soil layers
Wetlands 69% (100 Tg)
